Page 202 - Soil and water contamination, 2nd edition

Systems and models 189

An essential aspect in environmental modelling is the choice of the number of spatial
model dimensions. The number of spatial dimensions (zero to three) chosen to describe
an environmental system is likewise based on the spatial variability of the key factors
governing the state of the system. If this spatial variation is insignificant or irrelevant, a
zero-dimensional model or lumped model may be chosen, in which the system is spatially
averaged. Examples of widely used zero-dimensional models are so-called box models
or Mackay models that describe the average partitioning of pollutants among various
environmental compartments such as sediment, pore water, surface water, and various
trophic level s of the food chain (Mackay, 1991). Such models are commonly used in
environmental risk assessment; they are applied to larger areas, such as an entire lake or river
basin. In contrast, a distributed model takes account of the spatial variation of the patterns
of environmental entities and processes. One-dimensional models are generally used when
the spatial variation predominates in one direction: for example, the vertical distribution of
contaminants over a soil profile, or the longitudinal variation of water quality in a river. For
some river water quality problems a two-dimensional model may be required: for instance,
when modelling the dispersal of contaminants over part of a local floodplain. For more
complex surface water bodies a three-dimensional model may be needed: for instance, when
stratification of temperature in deep lakes or of salt concentration in estuaries occurs. Because
groundwater composition usually varies both vertically and laterally, groundwater quality
models usually also require three dimensions.
In recent years, there has been a growing interest in the development of spatially
distributed model s that simulate or predict the sources, transport, and fate of contaminants
at the catchment scale (see e.g. Addiscott and Mirza, 1998; De Wit, 2001; Whelan and
Gandolfi, 2002; Mourad and Van der Perk, 2004; Håkanson, 2004; Heathwaite et al., 2005;
Destouni et al., 2010; Zanardo et al., 2012). Concurrently, Mackay models to model the
partitioning of pollutants over the environmental compartments have also been implemented
in a spatially distributed manner (e.g. Coulibaly et al., 2004; Warren et al., 2005). Most
of the model studies mentioned have used GIS , because it is advantageous and useful for
preparing model input from large spatial data sets on catchment characteristics (e.g. altitude,
soil type, land use, hydrology ) and for the post-processing of model output. Some models are
entirely GIS-based, which implies that these models operate wholly in a GIS environment
using an integrated programming language (Burrough, 1996). These developments have
opened the door for scientists who want to construct their own environmental models
tailored to their specific purposes and available data (see Karssenberg, 2002).

EXERCISES

1. Give examples of at least three feedback mechanisms that control the decay of organic
matter in soil. Indicate whether these feedback mechanisms are positive or negative.

2. What is the difference between a closed system and open system?

3. Why is the drainage basin a convenient unit for studying contaminant transport and fate?

3
4. The nitrate concentration in a lake with a volume of 100 000 m which does not have
an outflow is determined by the input via atmospheric deposition and losses due to
-1
denitrification. The concentration in the lake is 0.2 mg l . The denitrification rate is
given by:
-1
-1
-1
-1
Denitrification (mg l d ) = 0.1 (d ) × nitrate concentration (mg l )

10/1/2013 6:44:42 PM
Soil and Water.indd 201 10/1/2013 6:44:42 PM
Soil and Water.indd 201